Lukáš Hejtmánek

High-definition multimedia for multiparty low-latency interactive communication

We describe a high-quality collaborative environment that uses High-Definition (HD) video to achieve near realistic perception of a remote site. The capture part, consisting of a HD camera, Centaurus HD-SDI capture card, and UltraGrid software, produces a 1.5 Gbps UDP data stream of uncompressed HD video that is transferred over a 10GE network interface to the high-speed IP network. The HD video stream displaying uses either a software-based solution with color depth down-sampling and field de-interlacing, or another Centaurus card. Data distribution to individual participants of the videoconference is achieved using a user-controlled UDP packet reflector based on the Active Element idea. The viability of this system has been demonstrated at the iGrid 2005 conference for a three-way high quality videoconference among sites in the Czech Republic, Louisiana, and California.

Virtual microscope interface to high resolution histological images

The Hypertext atlas of Dermatopathology, the Atlas of Fetal and Neonatal Pathology and Hypertext atlas of Pathology (this one in Czech only) are available at http://www.muni.cz/atlases. These atlases offer many clinical, macroscopic and microscopic images, together with short introductory texts. Most of the images are annotated and arrows pointing to the important parts of the image can be activated.The Virtual Microscope interface is used for the access to the histological images obtained in high resolution using automated microscope and image stitching, possibly in more focusing planes. Parts of the image prepared in advance are downloaded on demand to save the memory of the users computer. The virtual microscope is programmed in JavaScript only, works in Firefox/Mozilla and MSIE browsers without need to install any additional software.

Distributed Data Storage with Strong Offline Access Support

In this paper, we propose a distributed data storage framework that supports unrestricted offline access. The system does not explicitly distinguish between connected and disconnected states. Its design is based on a lock-free distributed framework that avoids update conflicts through file versioning. The feasibility of this framework is confirmed by a proof-of- concept implementation. We also demonstrate that the proposed lock-free replica synchronization algorithm scales well. A future work will include also direct support for non-versioned files.

Distributed Negative Cycle Detection Algorithms

Publisher Summary This chapter discusses the distributed negative cycle detection algorithms. Several new parallel algorithms for the single source shortest paths (SSSP) and for the negative cycle detection (NCD) problems on directed graphs with real edge weights and given by adjacency list are developed, analyzed, and experimentally compared. The algorithms are to be performed on clusters of workstations that communicate via a message passing mechanism. The chapter provides and analyzes parallel algorithms for the general SSSP and NCD problems for graphs specified with adjacency lists. The algorithms are designed for networks of workstations where the input graph is distributed over individual workstations communicating via a message passing interface. Based on experiments, the chapter concludes that in situations where a priori information about the graph is given the best choice in the Subtree Disassembly algorithm. For specific applications other algorithms or their combinations can be more suitable, as demonstrated by Walk to Root, in case of application to model checking.

Secure and Fault Tolerant Distributed Framework with Mobility Support

In this paper, we propose an architecture of distributed data storage framework that incorporates fault tolerance, mobility support, and security. Main goal of our system is to provide equal opportunities for both connected and disconnected clients. Consequence is that mutual exclusion may not be involved. Data storage systems without mutual exclusion suffer from update and name conflicts. We avoid the update conflicts using immutable data storage. Mutable data is provided via either file versioning or Redo Logs. The name conflicts are automatically resolved without users guidance, the file names are automatically changed to non-conflicting names, the directories are represented implicitly and thus we avoid conflict names connected to directories. Because the file names may be changed by the system, each file version is assigned an immutable globally unique identifier using which the file version can be accessed. Security model is based on certificates and VOMS attributes. This system is suitable for use within Grid VOs and it also supports services provided simultaneously to different VOs. Our prototype implementation exhibits very favorable performance so that it could be used as robust, secure, highly reliable and high performance Storage Elements.